METHODS ARTICLEpublished: 10 June 2014

doi: 10.3389/fmicb.2014.00286

An optimized method for the extraction of bacterial mRNAfrom plant roots infected with Escherichia coli O157:H7Ashleigh Holmes1, Louise Birse1, Robert W. Jackson 2 and Nicola J. Holden1*

1 Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee, UK2 School of Biological Sciences, The University of Reading, Knight Building, Whiteknights, Reading, UK

Edited by:

Adam Schikora, IPAZ, JLU Giessen,Germany

Reviewed by:

Saul Burdman, The Hebrew Universityof Jerusalem, IsraelThamarai Schneiders, Queen’sUniversity Belfast, UK

*Correspondence:

Nicola J. Holden, Cell and MolecularSciences, The James Hutton Institute,Invergowrie, Dundee, DD2 5DA, UKe-mail: nicola.holden@hutton.ac.uk

Analysis of microbial gene expression during host colonization provides valuable informa-tion on the nature of interaction, beneficial or pathogenic, and the adaptive processesinvolved. Isolation of bacterial mRNA for in planta analysis can be challenging where hostnucleic acid may dominate the preparation, or inhibitory compounds affect downstreamanalysis, e.g., quantitative reverse transcriptase PCR (qPCR), microarray, or RNA-seq.The goal of this work was to optimize the isolation of bacterial mRNA of food-bornepathogens from living plants. Reported methods for recovery of phytopathogen-infectedplant material, using hot phenol extraction and high concentration of bacterial inoculation orlarge amounts of infected tissues, were found to be inappropriate for plant roots inoculatedwith Escherichia coli O157:H7. The bacterial RNA yields were too low and increased plantmaterial resulted in a dominance of plant RNA in the sample.To improve the yield of bacterialRNA and reduce the number of plants required, an optimized method was developed whichcombines bead beating with directed bacterial lysis using SDS and lysozyme. Inhibitoryplant compounds, such as phenolics and polysaccharides, were counteracted with theaddition of high-molecular-weight polyethylene glycol and hexadecyltrimethyl ammoniumbromide. The new method increased the total yield of bacterial mRNA substantially andallowed assessment of gene expression by qPCR. This method can be applied to otherbacterial species associated with plant roots, and also in the wider context of food safety.

Keywords: food-borne pathogens, lettuce, spinach, rhizosphere, mRNA isolation

INTRODUCTIONThe analysis of bacterial gene expression is important in the deter-mination of how adaptation to different environments developsand informs on the roles of different genes during this process.Human pathogens are now recognized to interact with plants anduse them as hosts, as a result of recent high-profile outbreaks fromcontaminated fruit and vegetables (Cooley et al., 2007; Buchholzet al., 2011). Adaptation of food-borne pathogens to secondaryhosts has opened up new areas of research and investigation. Cur-rent research suggests that the interaction of human pathogens ismore complex than previously perceived in that they can persistfor long periods of time (reviewed in Brandl, 2006; Holden et al.,2009; Barak and Schroeder, 2012) and invoke an immune responsefrom the plant (Thilmony et al., 2006; Schikora et al., 2008; Royet al., 2013). Furthermore, the interactions include quite specificand targeted recognition of the plant host cells by the bacteria(Rossez et al., 2013). Important questions remain as to how thebacteria adapt to secondary hosts, for which analysis of bacterialgene expression is fundamental.

Bacterial gene expression is typically assessed using quantitativereverse transcriptase PCR (qPCR) or microarray techniques, andmore recently RNA-seq (An et al., 2013), which require the isola-tion of high quality and quantity of mRNA. Bacterial mRNA has ashort half-life and is less stable than eukaryotic mRNA transcripts,which have capped and polyadenylated RNA tails. Therefore,appropriate measures need to be in place to capture mRNA tran-scripts that may be inherently unstable, but nevertheless important

for bacterial adaptation. In addition to challenges with bacterialmRNA stability, extractions from mixed samples bring additionalconsiderations, not least for inhibitory compounds that may affectdownstream analysis. Although there are a number of publishedtechniques for the isolation of total RNA from bacteria-infectedplant leaves (Schenk et al., 2008; Soto-Suarez et al., 2010; Finket al., 2012; Goudeau et al., 2013), samples can still be dominatedby plant RNA making it challenging to assess bacterial mRNA. Inpublished reports, leaves were typically infected via infiltration,introducing a high bacterial inoculum into the sample. In otherstudies, total RNA was extracted from inoculated plant extracts(Mark et al., 2005; Hernandez-Morales et al., 2009; Kyle et al., 2010;Shidore et al., 2012), such as leaf lysates, which helped to reducethe dominance of plant RNA in the sample.

There are limited studies where bacterial expression has beenassessed directly from the plant root system (roots and rhizo-sphere). However, roots are known to support relatively highdensities of bacteria, providing a more favorable habitat thanthe phyllosphere. In general, published reports describe that ahigh number of plant roots is required to retrieve sufficient bacte-rial RNA (Matilla et al., 2007; Hou et al., 2012, 2013; Zysko et al.,2012). Further to this, the application of techniques describedfor RNA purification from infected leaves cannot always be suc-cessfully applied to roots. Therefore, optimization of the RNAextraction methods is required to obtain sufficient quantity andquality of bacterial mRNA, coupled with a reduction in the amountof accompanying plant root RNA. We optimized the method for

www.frontiersin.org June 2014 | Volume 5 | Article 286 | 1

Holmes et al. E. coli O157:H7 RNA extraction from plants

the food-borne pathogen Escherichia coli O157:H7, frequentlyassociated food-borne outbreaks from consumption of contam-inated spinach and lettuce. Although roots of these plants arenot consumed, the pathogen can colonize this habitat successfully(Wright et al., 2013), from where it can contaminate the edibleportion, either directly or through cross-contamination duringprocessing.

MATERIALS AND METHODSBACTERIAL STRAINS AND GROWTH CONDITIONSE. coli O157:H7 isolate Sakai stx-negative (Hayashi et al., 2001) wasroutinely cultured overnight in Luria broth at 37◦C, with aeration.For plant-infection assays, the bacteria were sub-cultured at 1:50dilution into 15 ml rich-defined MOPs media supplemented with0.2% glucose (RD MOPS glucose; Neidhardt et al., 1974), in a200 ml Erlenmeyer flask and incubated at 18◦C, with aeration for∼18 h. Bacterial cultures were diluted to OD600 of 0.2 (equivalentto ∼2 × 108 cfu ml−1) in sterile phosphate buffered saline (PBS)prior to infecting plant roots.

PLANT PROPAGATION AND INFECTIONLettuce (Lactuca sativa) cultivar All Year Round or spinach(Spinacia oleracea) cultivar Amazon seeds (Sutton Seeds, UK)were soaked in sterile distilled water for 2 h before being sur-face sterilized in 2% calcium hypochlorite solution (10 ml) for10 min. The seeds were then washed vigorously six times withsterile distilled water and germinated on distilled water agar(0.5% w/v) in the dark for 3–5 days, at ∼22◦C. Seedlings weretransplanted into 175 ml hydroponic tubes (Greiner, Fricken-hausen, Germany) containing autoclaved perlite and sterile 0.5×Murashige and Skoog (MS) medium (Sigma Aldrich, USA).Seedlings were grown in a cabinet with a light intensity of150 μmol m2 s−1 (16 h photoperiod) for a further 21 days at22◦C. To assess bacterial numbers from infected plant material,the roots were washed with PBS to remove excess, non-adherentbacteria. The root sample was homogenized using mortar andpestle and serially diluted for viable counts on selective agarplates.

RNA EXTRACTION AND PURIFICATIONTotal RNA was extracted from infected root samples either by theBPEX method (Schenk et al., 2008) or by the Bead/SDS/phenolmethod (Figure 1): a method which was adapted and optimizedfrom Jahn et al. (2008) and Schenk et al. (2008).

Sample preparationSample preparation was common to both methods. Whole plantswere gently removed from the hydroponic tubes and washed withsterile PBS to remove as much perlite as possible before bacte-rial infection. Plants were pooled into groups of 5 and the rootswere submerged into 20–25 ml bacterial suspension (OD600 = 0.2equivalent to ∼ 5 × 107–1 × 108 cfu ml−1), and incubated at 18◦Cfor 2 h. After incubation, the plants were removed from the bacte-rial suspension and all the tissue above the crown (shoots, leaves,and petioles) were aseptically removed with a sterile scalpel andthe roots immediately immersed into 20 ml ice-cold, 95% ethanol:5% phenol (pH 4.0). The sample was incubated on ice for 5 minand gently agitated on a vortex mixer to remove any excess and

loosely attached bacteria. The root sample was placed into a foilpacket and immediately stored in liquid nitrogen until all sampleshad been processed.

BPEX-Schenk method for extractionThe BPEX method was originally developed to recover Pseu-domonas syringae phytopathogen mRNA from infected plantmaterial (Schenk et al., 2008), and tested for recovery of E. colimRNA from infected roots. The main change from the publishedmethod was in the tissue type (root vs leaf disks), although theprotocol is provided to allow a direct comparison with the opti-mized method. Infected plant tissue was ground to a fine powderin liquid nitrogen with a pre-cooled pestle and mortar. The sam-ple (∼1 g) was transferred to an Eppendorf containing 750 μl ofbacteria plant extraction (BPEX) buffer [0.35 M glycine, 0.7 MNaCl, 2% (w/v) polyethylene glycol (PEG) 20000, 40 mM EDTA,50 mM NaOH, 4% (w/v) SDS] supplemented with 100 mM β-mercaptoethanol just prior to use. The sample was mixed on avortex mixer in the buffer prior to incubation at 95◦C for 90 s withshaking. An equal volume (750 μl) of phenol/chloroform mix (5:1,pH 4.0) (Sigma, St. Louis, USA) was added and the sample shakenfor 5 min to form an emulsion before centrifugation at 16,000 × gfor 7 min. The upper phase was collected and added to an equalvolume of phenol/chloroform mix (5:1, pH 4.0), shaken, and cen-trifuged again as the previous step. The upper phase was collectedand added to an equal volume of phenol/chloroform/isoamylalcohol (IAA) mix (25:24:1, pH 4.0) (Sigma, St. Louis, USA)shaken and centrifuged again as before. A volume of 495 μl ofthe upper phase was added to 550 μl chloroform/IAA mix (24:1)and overlayed with 55 μl pre-warmed (55◦C) hexadecyltrimethylammonium bromide (CTAB)/NaCl (10% CTAB, 0.7 M NaCl)solution. The suspension was shaken and centrifuged as before.The upper phase was added to 1/4 vol 8 M LiCl solution, mixed byinversion and the RNA precipitated at –20◦C for 30 min. To collectthe RNA, the sample was centrifuged at 16,000 × g for 20 min at4◦C and the RNA pellet resuspended in 100 μl RNase-free water.The RNA was cleaned and DNase treated using RNeasy Plant Minikit (Qiagen, Limburg, Netherlands) as per the manufacturer’sguidelines.

Bead/SDS/phenol method for extractionThe optimized method is represented in Figure 1. The sampleswere removed from the liquid nitrogen, beaten with a spatulato break up the root into smaller pieces, and transferred intoan 1.5 ml micro-centrifuge tube (DNase, RNase free: Ambion,Austin, USA) preloaded with ∼250 mg mixture of 1 mm glassand 0.1 mm silica beads (ThermoScientific Ltd., Waltham, USA),from which the weight was determined. Cooled, sterile equip-ment was used throughout the process. The micro-centrifugetube was then returned to liquid nitrogen for processing subse-quent samples until all of the samples were collected. For thelysis step, 800 μl of Tris-EDTA (TE) buffer (10 mM Tris-HCl;1 mM EDTA) supplemented with 500 μg ml−1 lysozyme, 0.1%SDS, and 100 mM β-mercaptoethanol was added to each rootsample. The tubes were placed into TissueLyser (Qiagen, Lim-burg, Netherlands) and agitated for 30 s with a 30 s intervalon ice for three cycles. After the last cycle, the samples were

Frontiers in Microbiology | Plant-Microbe Interaction June 2014 | Volume 5 | Article 286 | 2

Holmes et al. E. coli O157:H7 RNA extraction from plants

FIGURE 1 | Bead/SDS/phenol method of RNA extraction from infected

plant roots. A schematic of the optimized method (described in detail withinSection “Material and Methods”). The final concentrations of reagents are

indicated in the figure. Colored text has been used to highlight differentstages and treatments (e.g., blue for addition of beads/cold; green for lysis;red for shaking/heat).

returned to ice before transfer to a heatblock set at 64◦ C for2 min. To extract nucleic acids, the supernatant was collectedand pooled after two centrifugation steps: first at 100 × g for1 min to pellet the beads and large fragments of root, followedby a second at 8,600 × g for 2 min to compact the debris fur-ther, yielding more supernatant. One hundred millimolar sodiumacetate (pH 5.2) and 2% (w/v) PEG 20000 was added to thesupernatant and inverted to mix. An equal volume (∼1 ml) ofphenol (pH 4.0) was then added, mixed by inversion, and thesample incubated at 64◦C for 6 min, with the tubes invertedevery 40 s. The sample was transferred to an ice bath for 2 minbefore centrifugation at 16,000 × g for 10 min at 4◦C. The upperaqueous layer was added to an equal volume of chloroform/IAAmix (24:1) and 1/20 volume of pre-warmed (55◦C) CTAB/NaClsolution in a fresh micro-centrifuge tube. The sample was mixedby inversion and then centrifuged at 16,000 × g for 5 min at4◦C. The upper aqueous layer was added to 1/4 vol 8 M LiCl,

mixed by inversion, and then incubated at –20◦C for 20 min toovernight to precipitate the nucleic acid. The nucleic acid wasrecovered by centrifugation at 16,000 × g for 30 min at 4◦C.The pellet was resuspended in 100 μl RNase-free water and thesample cleaned and DNase treated using the RNeasy Plant Minikit (Qiagen, Limburg, Netherlands) as per the manufacturer’sinstructions.

Total RNA concentration was determined using a Nan-oDrop (Wilmington, USA) spectrophotometer and the relativeproportions of ribosomal RNA determined using a BioAnal-yser 2100 (Agilent Technologies, Santa Clara, USA), for bothmethods.

cDNA SYNTHESIS AND qPCR CONDITIONScDNA was transcribed from 1 μg total RNA using Superscript II(Invitrogen, Carlsbad, USA) following the random primer proto-col. A mixture of ten 11-mer oligonucleotide primers (Ea1 ± Ea10)

www.frontiersin.org June 2014 | Volume 5 | Article 286 | 3

Holmes et al. E. coli O157:H7 RNA extraction from plants

at 100 nM (Fislage et al., 1997) designed specifically for Enter-obacteriaceae mRNA was used. Quantitative reverse transcriptasePCR was carried out using specific primers for E. coli O157:H7gyrB housekeeping gene (gyrB.RT.F = CATCAGAGAGGTCG-GCTTCC; gyrB.RT.R = CATGGAGCGTCGTTATCCGA) usingStepOnePlusTM real-time PCR system (Applied Biosystems, LifeTechnologies, Carlsbad, USA) and iTaqTM Universal SYBR® GreenSupermix (Bio-Rad, Hercules, USA). Each 20 μl reaction vol-ume was composed of iTaq SuperMix, 300 nM of forward andreverse primers and 1 μl of cDNA (diluted 1:4). The PCR pro-gram consisted of an initial denaturation at 95◦C for 10 min,followed by 40 cycles of denaturation at 95◦C for 15 s, and anneal-ing and extension at 60◦C for 1 min. Melt curve analysis wasalso performed with an initial denaturation at 95◦C for 15 s, fol-lowed by annealing at 60◦C for 1 min with an 0.3◦C increase(step and hold) and final step of 95◦C for 15 s. Data were col-lected from three technical replicates and from two biologicalreplicates.

RESULTSEXTRACTION OF TOTAL RNA FROM PLANT ROOTS INFECTED WITHBACTERIAAn optimized method was previously reported for the extractionof P. syringae mRNA from infiltrated plant leaves (Schenk et al.,2008). A buffer system was developed that yielded 75–125 μg oftotal RNA per 150–200 mg sample, from the equivalent of 20infiltrated leaf disks (7 mm). We tested this protocol (termed the“BPEX” method) to determine if it could also be used to recovermRNA of a food-borne pathogen from infected roots. Our workfocuses on the bacterial genes induced during the stages of colo-nization of lettuce and spinach roots, and as such the roots in ourexperiments are not infected via infiltration; rather the bacteriacolonize the outer surface. This necessitated processing the entireroot to recover sufficient RNA. Also, to obtain equivalent bacterialnumbers to those reported, where each leaf disk was infiltratedwith 1 × 108 cfu bacteria, at least 20 root samples were pooledfor each extraction. In this method, the samples were groundto a fine powder in liquid nitrogen, prior to incubation in anextraction buffer (BPEX) coupled with cell lysis. RNA was puri-fied using traditional acidic phenol extraction and LiCl-mediatedprecipitation.

Use of the BPEX method typically yielded 3.6–9.6 μg of totalRNA from 250 mg E. coli O157:H7-infected lettuce roots. Anal-ysis of the total RNA showed strong bands corresponding with18S and 28S rRNA of lettuce (Figure 2A: Lanes 1, 2, 3) with onlyvery faint bacterial RNA corresponding to 16S and 23S rRNA inmixed samples (Figure 2A; Lanes 1, 2, 3). Examination of the elec-tropherogram trace showed that the bacterial rRNA peaks wereconsiderably smaller than the plant rRNA peaks, e.g., BPEX sample1 (Figure 2B). The pre-dominance of plant rRNA in the BPEX-prepared samples indicated that plant mRNA would also dominateover bacterial transcripts. Therefore, to increase the levels of bac-terial mRNA with a concomitant decrease in plant-derived RNA,a modified method was designed. The aim was threefold: (i) toreduce the number of plants required per extraction; (ii) to reducethe carry-over of plant tissue in the sample; and (iii) incorporatea lysis step which targeted bacterial cells.

FIGURE 2 |Total RNA from plant root extractions by BPEX and

Bead/SDS/phenol methods. Intact plant roots were suspended in 20 mlof bacterial inoculum (OD600 of 0.2) for 2 h at 18◦C. Total RNA wasextracted using either the BPEX or the Bead/SDS/phenol method and RNAquality was assessed by spectroscopy on a Bioanalyser 2100 machine(Agilent). (A) A montage of an electropherograph of the total RNA levelsfollowing extraction of inoculated plant material. The lanes show RNA froman in vitro culture of E. coli O157:H7 isolate Sakai only (lane S); uninfectedlettuce roots (lane L); E. coli O157:H7 Sakai infected lettuce roots (Lane 1,306 ng μl−1; Lane 4, 119 ng μl−1; Lane 5, 195 ng μl−1); and E. coliO157:H7 Sakai infected spinach roots (Lane 2, 248 ng μl−1; Lane 3,301 ng μl−1; Lane 6, 141 ng μl−1). Lanes 1, 2, and 3 show extraction usingthe BPEX method, and Lanes 4, 5, and 6 using the Bead method. Thesamples were run alongside a commercial RNA transcript ladder (0.2, 0.5,1.0, 2.0, 4.0, and 6.0 kb, Agilent); the 16S and 23S (bacterial), and 18S and28S (plant) rRNA bands are indicated. Electropherogram of E. coli O157:H7Sakai infected lettuce roots extracted with (B) the BPEX method (samplederived from lane 1) or with (C) the novel Bead/SDS/phenol method(sample derived from lane 4). The uninfected spinach control was similar tolettuce (lane L) (not shown).

Frontiers in Microbiology | Plant-Microbe Interaction June 2014 | Volume 5 | Article 286 | 4

Holmes et al. E. coli O157:H7 RNA extraction from plants

DEVELOPMENT OF THE BEAD/SDS/PHENOL PROTOCOLPhysical disruption of tissues and cells using inert beads has beenused for nucleic acid extraction from filamentous fungi (Leiteet al., 2012), microalgae (Kim et al., 2012), soil and sludge samples(Griffiths et al., 2000), and in commercial kits. We postulated thatthis method of disruption may also aid in bacterial RNA extrac-tion and used a combination of 1 mm glass beads and 0.1 mmsilica beads for sample lysis. Since tissue can become over-heatedduring agitation, which leads to nucleic acid degradation, the sam-ples were “rested” for 30 s on ice between the three treatmentcycles.

Bacterial cells were specifically targeted for lysis in an attemptto reduce plant-derived nucleic acid contamination, with theinclusion of lysozyme in the extraction buffer. However, a com-bination of the hot SDS/phenol extraction protocol describedpreviously (Jahn et al., 2008) with the bead beating step resultedin very low RNA yields; not greater than 0.5 μg. It was possiblethat contaminants or inhibitors from the plant material, such aspolysaccharides, phenolic compounds, and secondary metabolitesdeleteriously affected RNA recovery. Therefore, the protocol wasmodified to include additional steps after the initial lysis and theincorporation of PEG and CTAB/NaCl treatments, followed byLiCl precipitation (Figure 1). The inclusion of high-molecular-weight PEG in the extraction protocol promotes the removal ofpolysaccharides and phenolic compounds from plant tissues thatbind to or co-precipitate with RNA (Gehrig et al., 2000). CTAB is adetergent that acts to separate polysaccharides from nucleic acids(Chang et al., 1993; Jaakola et al., 2001). Lithium chloride is inef-ficient at precipitating DNA, proteins, or carbohydrates, and thusimproves RNA yield and purity compared to other nucleic acidprecipitation methods, i.e., ethanol and sodium acetate (Barlowet al., 1963; Cathala et al., 1983).

ANALYSIS OF RNA YIELDSA combination of physical disruption using beads together withthe chemical treatments (Figure 1) resulted in total RNA con-centrations that averaged 5.1 μg (range from 1.5 to 8.9 μg) from250 mg of infected root tissue. Assessment of the RNA showeda substantial increase in bacterial-derived 16S and 23S rRNA inbead-treated samples compared to BPEX samples (Figure 2A –compare lanes 4–6 with lanes 1–3). The optimized methodincreased the bacterial rRNA to levels that were equivalent or closeto plant rRNA (Figure 2C).

QUANTITATIVE REVERSE TRANSCRIPTASE PCR (qPCR) ANALYSISBacterial gene expression was quantified from the samplesobtained using the optimized method, by carrying out qPCR. ThegyrB gene was selected as a housekeeping gene that was expectedto be expressed under the conditions tested and expression fromthe root-derived samples was compared to samples obtained frombacteria grown under in vitro conditions. The amount of gyrBtranscript was found to be similar between both lettuce rootsand in vitro samples (Table 1), indicating similar copy numbersand stable gyrB expression under in vitro conditions and in plantextracts. It is of note that the amount of detectable gyrB was lowerin the infected spinach root extracts compared to lettuce roots(higher Ct value for gyrB).

Table 1 | Comparison of Ct values from E. coli O157:H7 isolate Sakai

for an in vitro culture and for infected plant roots.

Sample Average Ct ± SD

Sakai culture 23.435 ± 0.415

Sakai + lettuce roots 23.525 ± 0.295

Sakai + spinach roots 26.997 ± 0.527

In brief, plant roots were suspended in 20 ml bacterial inoculum (OD600 of 0.2)for 2 h at 18◦C, washed vigorously, and RNA prepared. The bacteria-only culturewas suspended in 20 ml 1× PBS for 2 h at 18◦C. The gyrB gene was quantifiedusing a StepOnePlus PCR machine (Applied Biosystems). The values show theaverage of six samples from two independent experiments.

QUANTIFICATION OF BACTERIATo quantify the bacteria present in the samples, the average numberof E. coli O157:H7 (Sakai) cells present in a 250 mg root sample(typically from 7 to 8 roots) was determined. The experimental set-up was as described for RNA extraction, although the roots werewashed with PBS instead of the initial step of mRNA preservationin the ice-cold phenol/ethanol mixture. On average, 5.6 × 107 cfuof E. coli O157:H7 (Sakai) were present in each lettuce sample. Thevalues for spinach roots were similar, which represent between 10and 15% of the initial inoculum used for infection. These figuresare equivalent to the limit of bacterial gene detection reported bySchenk et al. (2008) of ∼ 5 107 cfu/sample derived from 20 leafdisks.

DISCUSSIONThe isolation of total RNA from plant material is well defined forleaves, but less so for studies involving the roots or rhizosphere.The majority of reports focus on analysis of gene expressionfor phytopathogens and the number of examples describing theextraction of food-borne pathogen RNA from plants is limited.Here, we describe optimization of a method that resulted in high-quality bacterial mRNA from the infected roots of fresh produceplants: lettuce and spinach.

Reported investigations of gene expression of enteric bacteria,such as Salmonella enterica (Goudeau et al., 2013), E. coli K-12and E. coli O157:H7 (Kyle et al., 2010; Fink et al., 2012) used largequantities of plant material, e.g., 100 g of leaf tissue coupledwith high concentrations of bacteria, approx. 108 cfu per sam-ple. Samples were processed to separate bacteria from plant tissueusing physical methods, such as a Stomacher (Kyle et al., 2010;Goudeau et al., 2013) or by shaking (Fink et al., 2012) followedby filtration to remove plant debris. RNA was prepared eitherusing commercial kits (Kyle et al., 2010; Goudeau et al., 2013) ora hot-phenol method (Fink et al., 2012). Some of these studiesinvestigated gene expression on post-harvest material, using pre-packaged commercial leaves rather than propagating plants, whereit is possible to obtain the large amount of material required. Incontrast, investigation of bacteria gene expression on living plantroots presents challenges in obtaining similar weights of material.A hydroponic system for high-throughput propagation of lettuceplants was described recently (Hou et al., 2012, 2013), where ster-ilized seeds were germinated in 96-well pipette tip boxes. Wholetranscriptome analysis of E. coli was examined three days after

www.frontiersin.org June 2014 | Volume 5 | Article 286 | 5

Holmes et al. E. coli O157:H7 RNA extraction from plants

inoculation with the plant roots, although the report does notstate the weight of material required for RNA extraction. Themain difference in our method was the amount of plant mate-rial required and the time at which the samples were taken. Forexample, our method allows for gene expression analysis after ashort time of interaction, rather than after several days of colo-nization where the bacterial numbers are likely to have increased,dependent on the inoculation time and plant-bacterial systeminvestigated.

The optimized method uses a combination of bead-beating,SDS lysis, phenol extraction, and CTAB purification to extracthigh-quality bacterial RNA from as little as 250–300 mg infectedroots. Optimization has been carried out for two fresh produceplant species that have previously been associated with large-scaleoutbreaks of food-borne pathogens. It was interesting to note thatthe efficiency of recovery was lower for spinach root compared tolettuce, although the numbers of bacteria recovered from the planttissue were similar. Since E. coli O157:H7 Sakai adheres to lettuceand spinach roots in comparable numbers (Wright et al., 2013), weanticipated similar levels of gyrB transcript. The reduction in levelmay be as a result of plant-associated inhibitors in spinach roots,or because of different plant-derived environmental cues acting ongyrB expression. Therefore, we suggest that the method needs to bevalidated if it is adopted for other plant species. Verification of thetechnique using qPCR shows that this method could be appliedto food safety settings, e.g., for detection of food-borne pathogensin fresh produce. It may be possible to couple this technique tostandard methods, which would enable examination of viabilityand expression of genes of interest, e.g., toxin genes. Furthermore,it facilitates analysis of bacterial gene expression in planta for notonly food-borne pathogens but also other plant-associated bacte-ria, providing an insight into the adaptive processes that underpinhost–microbe interactions.

AUTHOR CONTRIBUTIONSAl authors were involved in the design of the work and interpre-tation of data; Ashleigh Holmes and Louise Birse were involved indata acquisition and analysis; Ashleigh Holmes, Louise Birse, andNicola J. Holden drafted the work; all authors approved the finalversion.

ACKNOWLEDGMENTSThe work was supported by BBSRC grant BB/I014179/1 and Scot-tish Government RESAS funding. The authors thank Pete Headleyand Jenny Morris (James Hutton Institute) for technical assistance;Sonia Humphris and the reviewers for helpful suggestions.

REFERENCESAn, S.-Q., Febrer, M., Mccarthy, Y., Tang, D.-J., Clissold, L., Kaithakottil, G.,

et al. (2013). High-resolution transcriptional analysis of the regulatory influ-ence of cell-to-cell signalling reveals novel genes that contribute to Xanthomonasphytopathogenesis. Mol. Microbiol. 88, 1058–1069. doi: 10.1111/mmi.12229

Barak, J. D., and Schroeder, B. K. (2012). Interrelationships of food safety and plantpathology: the life cycle of human pathogens on plants. Annu. Rev. Phytopathol.50, 241–266. doi: 10.1146/annurev-phyto-081211-172936

Barlow, J. J., Mathias, A. P., Williamson, R., and Gammack, D. B. (1963). Asimple method for the quantitative isolation of undegraded high molecularweight ribonucleic acid. Biochem. Biophys. Res. Commun. 13, 61–66. doi:10.1016/0006-291X(63)90163-3

Brandl, M. T. (2006). Fitness of human enteric pathogens on plants andimplications for food safety. Annu. Rev. Phytopathol. 44, 367–392. doi:10.1146/annurev.phyto.44.070505.143359

Buchholz, U., Bernard, H., Werber, D., Bohmer, M. M., Remschmidt, C., Wilking,H., et al. (2011). German outbreak of Escherichia coli O104:H4 associated withsprouts. N. Engl. J. Med. 365, 1763–1770. doi: 10.1056/NEJMoa1106482

Cathala, G., Savouret, J. F., Mendez, B., West, B. L., Karin, M., Martial, J. A., et al.(1983). A method for isolation of intact, translationally active ribonucleic acid.DNA 2, 329–335. doi: 10.1089/dna.1983.2.329

Chang, S., Puryear, J., and Cairney, J. (1993). A simple and efficient methodfor isolating RNA from pine trees. Plant Mol. Biol. Rep. 11, 113–116. doi:10.1007/BF02670468

Cooley, M., Carychao, D., Crawford-Miksza, L., Jay, M. T., Myers, C., Rose,C., et al. (2007). Incidence and tracking of Escherichia coli O157:H7 in amajor produce production region in California. PLoS ONE 2:e1159. doi:10.1371/journal.pone.0001159

Fink, R. C., Black, E. P., Hou, Z., Sugawara, M., Sadowsky, M. J., and Diez-Gonzalez, F. (2012). Transcriptional responses of Escherichia coli K-12 andO157:H7 associated with lettuce leaves. Appl. Environ. Microbiol. 78, 1752–1764.doi: 10.1128/aem.07454-11

Fislage, R., Berceanu, M., Humboldt, Y., Wendt, M., and Oberender, H. (1997).Primer design for a prokaryotic differential display RT-PCR. Nucl. Acids Res. 25,1830–1835. doi: 10.1093/nar/25.9.1830

Gehrig, H. H., Winter, K., Cushman, J., Borland, A., and Taybi, T. (2000). Animproved RNA isolation method for succulent plant species rich in polyphenolsand polysaccharides. Plant Mol. Biol. Rep. 18, 369–376. doi: 10.1007/BF02825065

Goudeau, D. M., Parker, C. T., Zhou, Y., Sela, S., Kroupitski, Y., and Brandl, M.T. (2013). The Salmonella transcriptome in lettuce and cilantro soft rot revealsa niche overlap with the animal host intestine. Appl. Environ. Microbiol. 79,250–262. doi: 10.1128/aem.02290-12

Griffiths, R. I., Whiteley, A. S., O’Donnell, A. G., and Bailey, M. J. (2000). Rapidmethod for coextraction of DNA and RNA from natural environments for anal-ysis of ribosomal DNA- and rRNA-based microbial community composition.Appl. Environ. Microbiol. 66, 5488–5491. doi: 10.1128/AEM.66.12.5488-5491.2000

Hayashi, T., Makino, K., Ohnishi, M., Kurokawa, K., Ishii, K., Yokoyama, K.,et al. (2001). Complete genome sequence of enterohemorrhagic Escherichia coliO157:H7 and genomic comparison with a laboratory strain K-12. DNA Res. 8,11–22. doi: 10.1093/dnares/8.1.11

Hernandez-Morales, A., De La Torre-Zavala, S., Ibarra-Laclette, E., Hernandez-Flores, J., Jofre-Garfias, A., Martinez-Antonio, A., et al. (2009). Transcriptionalprofile of Pseudomonas syringae pv. phaseolicola NPS3121 in response to tissueextracts from a susceptible Phaseolus vulgaris L. cultivar. BMC Microbiol. 9:257.doi: 10.1186/1471-2180-9-257

Holden, N., Pritchard, L., and Toth, I. (2009). Colonization outwith the colon: plantsas an alternative environmental reservoir for human pathogenic enterobacteria.FEMS Microbiol. Rev. 33, 689–703. doi: 10.1111/j.1574-6976.2008.00153.x

Hou, Z., Fink, R. C., Black, E., Sugawara, M., Zhang, Z., Diez-Gonzalez, F.,et al. (2012). Gene expression profiling of Escherichia coli in response to inter-actions with the lettuce rhizosphere. J. Appl. Microbiol. 113, 1076–1086. doi:10.1111/j.1365-2672.2012.05412.x

Hou, Z., Fink, R. C., Sugawara, M., Diez-Gonzalez, F., and Sadowsky, M. J. (2013).Transcriptional and functional responses of Escherichia coli O157:H7 growingin the lettuce rhizoplane. Food Microbiol. 35, 136–142. doi: 10.1016/j.fm.2013.03.002

Jaakola, L., Pirttila, A. M., Halonen, M., and Hohtola, A. (2001). Isolation of highquality RNA from bilberry (Vaccinium myrtillus L.) fruit. Mol. Biotechnol. 19,201–203. doi: 10.1385/mb:19:2:201

Jahn, C. E., Charkowski, A. O., and Willis, D. K. (2008). Evaluation of isolationmethods and RNA integrity for bacterial RNA quantitation. J. Microbiol. Methods75, 318–324. doi: 10.1016/j.mimet.2008.07.004

Kim, B.-H., Ramanan, R., Cho, D.-H., Choi, G.-G., La, H.-J., Ahn, C.-Y., et al.(2012). Simple, rapid and cost-effective method for high quality nucleic acidsextraction from different strains of Botryococcus braunii. PLoS ONE 7:e37770.doi: 10.1371/journal.pone.0037770

Kyle, J. L., Parker, C. T., Goudeau, D., and Brandl, M. T. (2010). Transcriptomeanalysis of Escherichia coli O157:H7 exposed to lysates of lettuce leaves. Appl.Environ. Microbiol. 76, 1375–1387. doi: 10.1128/AEM.02461-09

Frontiers in Microbiology | Plant-Microbe Interaction June 2014 | Volume 5 | Article 286 | 6

Holmes et al. E. coli O157:H7 RNA extraction from plants

Leite, G. M., Magan, N., and Medina, A. (2012). Comparison of differentbead-beating RNA extraction strategies: an optimized method for filamen-tous fungi. J. Microbiol. Methods 88, 413–418. doi: 10.1016/j.mimet.2012.01.011

Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H., Haynes, J., Baysse, C., et al.(2005). Transcriptome profiling of bacterial responses to root exudates identifiesgenes involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102,17454–17459. doi: 10.1073/pnas.0506407102

Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and Ramos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving forces ofbacterial life in the rhizosphere. Genome Biol. 8, R179. doi: 10.1186/gb-2007-8-9-r179

Neidhardt, F. C., Bloch, P. L., and Smith, D. F. (1974). Culture medium forenterobacteria. J. Bacteriol. 119, 736–747.

Rossez, Y., Holmes, A., Wolfson, E. B., Gally, D. L., Mahajan, A., Pedersen, H.L., et al. (2013). Flagella interact with ionic plant lipids to mediate adherenceof pathogenic Escherichia coli to fresh produce plants. Environ. Microbiol. doi:10.1111/1462-2920.12315 [Epub ahead of print].

Roy, D., Panchal, S., Rosa, B. A., and Melotto, M. (2013). Escherichia coli O157:H7induces stronger plant immunity than Salmonella enterica TyphimuriumSL1344. Phytopathology 103, 326–332. doi: 10.1094/PHYTO-09-12-0230-FI

Schenk, A., Weingart, H., and Ullrich, M. S. (2008). Extraction of high-qualitybacterial RNA from infected leaf tissue for bacterial in planta gene expressionanalysis by multiplexed fluorescent Northern hybridization. Mol. Plant Pathol. 9,227–235. doi: 10.1111/j.1364-3703.2007.00452.x

Schikora, A., Carreri, A., Charpentier, E., and Hirt, H. (2008). The dark side ofthe salad: Salmonella typhimurium overcomes the innate immune response ofArabidopsis thaliana and shows an endopathogenic lifestyle. PLoS ONE 3:e2279.doi: 10.1371/journal.pone.0002279

Shidore, T., Dinse, T., Öhrlein, J., Becker, A., and Reinhold-Hurek, B. (2012). Tran-scriptomic analysis of responses to exudates reveal genes required for rhizospherecompetence of the endophyte Azoarcus sp. strain BH72. Environ. Microbiol. 14,2775–2787. doi: 10.1111/j.1462-2920.2012.02777.x

Soto-Suarez, M., Bernal, D., Gonzalez, C., Szurek, B., Guyot, R., Tohme, J., et al.(2010). In planta gene expression analysis of Xanthomonas oryzae pathovaroryzae, African strain MAI1. BMC Microbiol. 10:170. doi: 10.1186/1471-2180-10-170

Thilmony, R., Underwood, W., and He, S. Y. (2006). Genome-wide transcriptionalanalysis of the Arabidopsis thaliana interaction with the plant pathogen Pseu-domonas syringae pv. tomato DC3000 and the human pathogen Escherichia coliO157:H7. Plant J. 46, 34–53. doi: 10.1111/j.1365-313X.2006.02725.x

Wright, K. M., Chapman, S., Mcgeachy, K., Humphris, S., Campbell, E., Toth,I. K., et al. (2013). The endophytic lifestyle of Escherichia coli O157:H7: quan-tification and internal localization in roots. Phytopathology 103, 333–340. doi:10.1094/PHYTO-08-12-0209-FI

Zysko, A., Sanguin, H., Hayes, A., Wardleworth, L., Zeef, L. H., Sim, A., et al. (2012).Transcriptional response of Pseudomonas aeruginosa to a phosphate-deficientLolium perenne rhizosphere. Plant Soil 359, 25–44. doi: 10.1007/s11104-011-1060-z

Conflict of Interest Statement: The authors declare that the research was conductedin the absence of any commercial or financial relationships that could be construedas a potential conflict of interest.

Received: 24 March 2014; paper pending published: 27 April 2014; accepted: 23 May2014; published online: 10 June 2014.Citation: Holmes A, Birse L, Jackson RW and Holden NJ (2014) An optimized methodfor the extraction of bacterial mRNA from plant roots infected with Escherichia coliO157:H7. Front. Microbiol. 5:286. doi: 10.3389/fmicb.2014.00286This article was submitted to Plant–Microbe Interaction, a section of the journalFrontiers in Microbiology.Copyright © 2014 Holmes, Birse, Jackson and Holden. This is an open-access articledistributed under the terms of the Creative Commons Attribution License (CC BY). Theuse, distribution or reproduction in other forums is permitted, provided the originalauthor(s) or licensor are credited and that the original publication in this journal is cited,in accordance with accepted academic practice. No use, distribution or reproduction ispermitted which does not comply with these terms.

www.frontiersin.org June 2014 | Volume 5 | Article 286 | 7